1. Field of the Invention
The present invention is used in optical communications, and relates to techniques for transferring optical burst data in a communication network used in transferring burst data using optical wavelengths as communication media.
This application is based on patent applications Nos. 2001-337898, 2001-337920, 2001-338058, and 2001-338067 filed in Japan, the content of which is incorporated herein by reference.
2. Description of the Related Art
In conventional transferring of IP (internet protocol) data in optical communications, a plurality of routers are provided between the start point and the end point, and the data is transferred via these routers. The routers determine the next routers which the data is to be transferred to by referring to IP addresses, appended to the packets.
In an optical communication network of this type, when reading the header information of the packet, the optical signal is temporarily converted to an electrical signal. The router determines the next router which the data is to be transferred to in compliance with the IP address, read in this manner.
When transferring data in bursts, it is inefficient to read each of the IP addresses of the packets one by one; therefore, a cut through path between the start point and end point is set in advance, and, since the IP headers need not be read in segments of this cut through path, the optical signal can be transferred at high speed.
In this type of conventional optical communication, when transferring burst data by using the cut through path, the cut through path must be established in advance. To establish the cut through path in advance, a request to establish the cut through path is sent to each of the routers from the start point to the end point, the preliminary establishment of the cut through path being completed when the routers have complied with this request.
The procedure for establishing the cut through path at the routers comprises (1) converting the header information of the optical IP packet which has arrived to an electrical signal, and (2) determining whether information corresponding to a request to establish the cut through path is contained in the header information. When the information contains an establishment request, a wavelength is selected for establishing the cut through path, and the cut through path is established when a wavelength has been selected. Then, since there is no longer any need to read the header information of the IP packet on the wavelength which the cut through path has been established at, the establishment is altered in order to differentiate this wavelength from other wavelengths, where the IP packet is transferred in the form of an optical signal.
However, this type of establishment procedure requires a considerable amount of time. For example, let us assume that 2 t seconds are needed to establish the cut through path. When it takes 3 t seconds to transfer burst data using this cut through path, a total of 5 t seconds is required to establish the cut through path and transfer the burst data. In this example, nearly half of the total time of 5 t seconds to transfer the burst data is taken up by establishing the cut through path. If the time taken to establish the cut through path beforehand could be used for transferring data, the transferring efficiency would be greatly increased.
Although a cut through path is most appropriate for transferring burst data using a wavelength path, the quality of the optical signal deteriorates during transmission due to various factors. Consequently, there are inevitable limitations on the length of a cut through path which signals can be transmitted along without 3R processing.
Conventionally, the places to be 3R processed are determined by calculation. This calculation has two problems. The first is that deterioration in signal quality has a variety of causes, so that a huge amount of data must be processed in order to consider the great number of causes in the calculation. This increases the time needed to complete the calculation. The second problem is that the result obtained by such a calculation may not be highly accurate, necessitating safety measures such as providing extra places for 3R processing; this is not an efficient use of network resources.